To me, it’s not clear from the documentation, if rasa automatically saves the trained model into a S3 bucket, or we should copy the trained model manually?
I tried to train the model using rasa connected to S3 model, I did everything according to documentation, but still, model is saved only locally.
Any ideas?
Update - it creates the bucket, but model is only saved locally. I’ve enabled S3 logs and only bucket creation is logged, nothing more. Any ideas?
Mhm, normally it should save the models there. Do you get any error with rasa --debug <your other parameters>?
Hi @shota, hi @Tobias_Wochinger any news about save model in s3 automatically … i can t do it 
@kchahid Can you please provide some context and what the problem is?
so my probleme is : i wanna save model rasa automatically in s3 after the training like : rasa train --out [BUCKET S3] (i dont do that is only to explain)
i have set credentials in ~/.aws/ and in the config.yml (rasa config)
[…]
bucket_name: “XXX”
aws_region: “eu-west-1”
storage: “aws”
[…]
but i m not enbale to save model directly in s3
@Tobias_Wochinger
Same question here. The docs don’t seem to mention training the model and storing to remote storage. They focus on fetching and loading the model.